Iterative reconstruction methods in magnetic resonance elastography (MRE) inversion can allow incorporating sparsity prior and can provide displacement filtering. However the problem is difficult to converge as MRE inversion is a non-convex and ill-conditioned problem. Here, we presented 3D ERBA with total variation prior on the elasticity, which showed good convergence by utilizing a bi-convex optimization of 3D finite element modeling of elastic wave. Extensive experiments on numerical phantom, agar phantom, and in-vivo liver study showed promising indications in detecting inclusion and abnormality, and improving the diagnostic efficacy of MRE.
This abstract and the presentation materials are available to members only; a login is required.